# The Working Sequence

> From *The Decision Product* (Appendix A). Free to use and adapt.
>
> This is the order the book argues for, collected in one place so you can
> find it on a Monday. It is a checklist, not a methodology. Nobody should ever
> see it but you, it does not go in the write-up, and a decision is not better
> for having visibly passed through every step.
>
> Two things before the list. The steps are numbered because they have an
> order, not because you pass through each one once: the middle of the list is
> a loop, and you should expect to run it several times. A candidate that dies
> in week two is a successful week two.
>
> And the sequence is not a budget. Most decisions do not deserve every step.
> The skill actually being taught is knowing which ones the decision in front
> of you can skip.

---

**1. Name the choice as a sentence.**
The person who can say yes, the date by which they must, and whether you are
setting a one-shot commitment or a rule that will run repeatedly. If you
cannot write this sentence, nothing downstream will fix it.

**2. Invent the menu.**
The alternatives from the email plus the status quo are a first draft. Add at
least one commitment nobody in the original request had words for. Screen with
dominance, and with what the owner can actually do this cycle. **Activities are
not alternatives.**

**3. Write down the candidate ranking.**
What you would pick this afternoon if forced, and why you suspect it. It is
allowed to be wrong. It is not allowed to be blank or unspoken.

**4. Cut the question into parts, then prune.**
Draw the tree, check that the branches neither overlap nor leave a gap, and
mark every branch that cannot change the ranking. Write the sentence explaining
why you are leaving those alone; you will need it later.

**5. Write the claims that would kill the candidate.**
Two to four, of the same kind, each one a statement a result could falsify.
Decide now what result would falsify each, **before any query runs**.

**6. Choose the cheapest test that could change your mind, and run it.**
Not the most thorough test, and not the most interesting method. Complexity is
added later, and only when a claim requires it.

**7. Put the consequences in units that bite.**
Dollars where dollars are honest, stranded jobs where they are not, and an
explicit exchange rate wherever two units have to be traded against each other.
House metrics are not consequences.

**8. Name what would flip the ranking.**
The specific quantity, the specific threshold, and what it would cost to find
out. Include the cost of undoing the choice if the alternatives differ there.

**9. Ask who undoes this.**
Find the adjacent team that will dislike the recommendation if it works as
designed, and go ask them what they will do about it. If the answer is that
they will work around it, you have not finished.

**10. State the diagnosis, then the recommendation.**
A mechanism you could intervene on, and then a commitment among the named
alternatives. If you cannot name a mechanism, say that you are picking without
one rather than dressing the guess up.

**11. Write the headings before the prose.**
Full sentences, each an assertion. Read them in sequence and check that they
make the case alone. Rearrange until they do, then write underneath them.

**12. Write each party's no-deal.**
What they will do if this meeting produces nothing. If someone's walk-away
already beats your best alternative on the interest they actually hold, stop
arguing and change the menu, or stop.

**13. Write the five parts of the rule.**
The state it can see, the decision it may take, the information that will
arrive, how the next state is produced, and what you are accumulating. If a row
is blank, you have a slogan about a rule.

**14. Specify the rule so that it can run.**
Owner, cadence, thresholds, what happens when the inputs are missing, how an
exception is priced, what you will watch, and the date you revisit. The pieces
have to point the same way.

---

## If you have two days rather than six weeks

The sequence compresses to four steps and still works:

1. Name the choice.
2. Write the candidate ranking.
3. Run the single cheapest test that could kill it.
4. Write the note with the flip condition stated plainly.

What you lose is confidence. The honest response to that is to say so in the
write-up, rather than to pad the analysis until it looks like six weeks.

---

The Decision Product  ·  thedecisionproduct.com